Screaming in the Cloud - Episode 10: Education is Not Ready for Teacherless

Episode Date: May 16, 2018

Like migrating caribou, you tend to follow the trends of what clients are doing, which dictates what you work on as a consultant. Today, we’re talking to Lynn Langit, an independent Cloud architect. She is an AWS Community Hero, Google Cloud developer expert, and former Microsoft MVP. Lynn is a lifelong learner, and she has worked broad and deep across all three large providers. These days, she works mostly with Google Cloud and AWS, rather than Azure, because that’s what her clients are using. Some of the highlights of the show include: Differences between the West Coast and global use of Cloud Education is key; Lynn is th co-founder of Teachingkidsprogramming.org Lynn helped create curriculum and resources for school-age children; even her young daughter taught classes on how to code Training for teachers was also needed, so TKP Labs was formed to offer fee-based teacher and developer training Lynn started with classroom training, but has transitioned to online learning Lynn is focusing on Big Data projects and using tools to solve real-world problems Pre-processing and batching data, but not streaming it AWS, Azure, and Google Cloud are all coming out with Big Data-oriented tools Companies need to understand when the market is ready to accept a new paradigm; in the data world, change is more slow than in the programming world If you touch a database and get burned, you are not willing to use it again; or you may have never tried to archive your data; hire a consultant to help you Machine learning APIs give customers value quickly; review them before building custom models Migrating data can be a costly project and restricts where the data lives As Cloud proliferates, how will that impact technical education? Lynn’s Cloud for College Students to the rescue! Shift from interactive to unidirectional, one-to-many learning styles; the Cloud is ready for serverless, but education is not ready for teacherless Road that many of us walked to get to technical skills no longer exists; how to become a modern technologist Ageism: By age 40, you are considered a manager or useless; don’t be afraid to learn something new Links: Digital Ocean AWS Community Hero Microsoft Azure Teachingkidsprogramming.org Digigirlz TKP Labs Lynn Langit on Lynda.com Commonwealth Scientific and Industrial Research Organisation Google BigQuery Amazon Athena AWS Glue Cloud Dataflow Cloud Dataprep Lambda Amazon EC2 Learn Python the Hard Way .

Hello, and welcome to Screaming in the Cloud, with your host, cloud economist Corey Quinn. This weekly show features conversations with people doing interesting work in the world of cloud, thoughtful commentary on the state of the technical world, and ridiculous titles for which Corey refuses to apologize. This is Screaming in the Cloud. This week's episode of Screaming in the Cloud is generously sponsored by DigitalOcean. I would argue that every cloud platform out there biases for different things. Some bias for having every feature you could possibly want offered as a managed service at

varying degrees of maturity. Others bias for, hey, we heard there's some money to be made in the cloud space. Can you give us some of it? DigitalOcean biases for neither. To me, they optimize for simplicity. I polled some friends of mine who are avid DigitalOcean supporters about why they're using it for various things, and they all said more or less the same thing. Other offerings have a bunch of shenanigans, root access and IP addresses. DigitalOcean makes it all simple. In 60 seconds, you have root access to a Linux box with an IP. That's a direct quote, albeit with profanity about other providers taken out. DigitalOcean also offers fixed price offerings. You always know what you're going to wind up paying this month,

so you don't wind up having a minor heart issue when the bill comes in. Their services are also understandable without spending three months going to cloud school. You don't have to worry about going very deep to understand what you're doing. It's click button or make an API call and you receive a cloud resource. They also include very understandable monitoring and alerting. And lastly, they're not exactly what I would call small time. Over 150,000 businesses are using them today. So go ahead and give them a try. Visit do.co slash screaming, and they'll give you a free $100 credit to try it out.

That's do.co slash screaming. Thanks again to DigitalOcean for their support of Screaming in the Cloud. Welcome to Screaming in the Cloud. I'm Corey Quinn. Joining me today is Lynn Langit, who is an independent cloud architect and, among other things, holds the somewhat dubious honor of being an AWS community hero, a Google Cloud developer expert, and formerly a Microsoft MVP. I believe all at the same time. Welcome to the show. Thanks for having me. I'm excited to talk with you. So it's fascinating to sit down and talk to someone who is broad and deep across all three of the large public cloud providers. Usually someone specializes in one or at most two. Personally, I've gone very deep with AWS historically, but I cannot speak

authoritatively on too much past that. What is that like? So it's been interesting. I've been independent for seven years. And part of that, I actually was an employee of Microsoft for five years. And before that, I was a partner. So this past seven years have been a sea change for me. I mean, it's been really interesting. The beginning of it, I did a lot of work on Azure because I came out of the Microsoft ecosystem, but I did a lot of work on Amazon because they were the dominant cloud provider. Now, as the years go by, I am finding myself doing more work on Google Cloud and Amazon and less on Azure. Nothing negative about Azure. It's just what my customers are looking for. So it's interesting to see the way my consultancy has moved over time. It always seems that as a consultant myself,

that following the trends of what clients are doing dictates what I find myself working on. I can sit and say that a particular service or product or company's offering is terrific, and that is the thing I want to focus on. But if people aren't using it and aren't reaching out and asking for help around that particular thing, I find that I don't have much of a business. So I definitely understand what you're saying with respect to, I guess, following the trends almost like they're migrating caribou. Yeah. I live in Southern California and I do a lot of my work on the West Coast of the US, which is kind of a strange island in the global world of cloud because as a contrast to that, I also do work globally. So I either do work on the West Coast or somewhere else on Earth. And I find that these two

partitions are different. West Coast tends to embrace the new new and tends to be first up. Of course, we have wonderful connectivity, so that helps. I've worked in parts of the world where connectivity is not a given, and that really changes the game. One of the interesting things that I find that you've been focusing on has been the idea of education. As far as you're one of the co-founders of teachingkidsprogramming.org. You focus on getting people further along the path to technical mastery than they already are, which is incredibly valuable and incredibly important and is often an area that seems to get short shrift. How did you get involved in that? So I've been working with educating children around coding for 10 years.

It really started when I joined Microsoft in 2007. I joined a team of 62 people. I was a U.S. national technical presenter, and there were only two women on the team. And as a part of accepting that role at Microsoft, I actually negotiated, and I said that I would like to spend 25% of my time working on improving the pipeline, getting more people of color, women in technology. And that landed as me running a program called DigiGirls in Southern California, which were events for high school girls. And there wasn't any national curriculum at the time. So I worked with my then eight-year-old daughter and developed some curriculum.

And because there was a vacuum, that curriculum was used globally and got a lot of iterations and I learned quickly. Also for my job, I traveled globally. And so I kind of did a two-four all the time. I would go to the technical conference and speak about the shiny new Microsoft technology. And then I would hold a DigiGirls event.

So we got a lot of learning. Now, when I left Microsoft four or five years in, I discovered from working with my various communities that, particularly in the U.S., there was this situation of we had AP Java in high school, but there was no curriculum available to move the kids from visual learning like a scratch or a squeak to Java. So there was nothing in the middle school. So we took our work that we had done in small basic and we ported it to a middle school kid version of Java. And we also worked on the curriculum and the deliverables around the curriculum. So videos on how to use it, lesson plans, and worked with some middle school teachers. In fact, by then my daughter was in middle school. Kind of a funny story. She taught

a class of seventh graders when she was a seventh grader after school. And she also taught her history teacher how to code. And now she's for many years, subsequently, she's gone back and co-taught with him at our middle school. So it's kind of a cool story. It's fantastic watching people go from student to teacher. Something that I've always appreciated about conducting trainings is even when I take something that I thought I knew intimately and I teach it to someone else, I learned that, oh, my stars, I have no idea how this really works. And it's almost like going back to basics in a way that surprises me. Oh, that's for sure. So it's been really interesting. The organization is actually split

because it was labor of love on my part. And I self-funded most of the Java developers that I paid because I'm not a Java developer, ironically. So I paid Java developers to build out the curriculum with me. There's 80 lessons now, and it could be a full year, and it's being used by Google Analytics, about 16 states and 10 countries. So that's pretty exciting. But what was also needed was training for teachers, because the model that was really being adopted with the open source and free scenario that I had set up was teachers who were super achievers who were teaching themselves to code were using the curriculum. But there's a whole bunch of teachers

that really are wanting more training. They want to use their teacher and service days. So we spawned a child nonprofit called TKP Labs that does for fee teacher-based training. And we've had some really good success with that too. In fact, we're working with our first entire school district in Santa Barbara, California. And it's an interesting model. It's being sponsored by a local business and they're doing teacher training and developer training side-by-side. So we're really, really excited about that. This would be enough for a full-time job in and of itself from what you're describing, but you go beyond that.

You have a series of videos on lynda.com, which is now a LinkedIn property, which is now itself a Microsoft property and all roads lead to Rome. But on that site, you at last count had 17 courses available. By the time this publishes, you'll probably have five more. The rate things tend to move in this space. So is that an outgrowth of your work teaching kids then moving into adult education? Is this something that just sort of dovetailed along with it? Or did it come about from somewhere completely different? No, it's actually the other way. I was a classroom trainer. I'm middle-aged, so I've been training for a long time. So in the dot-com boom, I was a Microsoft certified trainer for eight years. And that's how I got into tech. I actually came out

of the business world and I took certification exams and then became a teacher, starting in the networking side and then really finding a love for databases. Ended up writing books about SQL Server and all that way back when, 2007, 2008. So I had this tech training background, this classroom training. Well, what has happened is classroom training has pretty much gone away because of online. And so I've really morphed that into online training. And I've trained for a number of different providers. I did some work for Pluralsight. I did some work for some other guys too, a long time ago. But the home I've kind of found is lynda.com, et cetera, whatever, Microsoft basically, because

they have a really neat setup. Well, first of all, they're in Santa Barbara, which is a wonderful place to go, but they have recording studios. And so you go up and if you listen to one of my courses, well, this is going to be professionally produced, so it'll probably sound the same. But I always marvel at how great they make me sound and how great they make me look. So it's a full team effort, which results in, I think, a better experience for students. So as a teacher, I'm happy to be a part of what they're doing up there. And it sounds like it's something that's absolutely fantastic.

Every person I've spoken to who's engaged with it in some way, shape, or form comes away either on the teacher side or the student side with rave reviews. Moving on to a slightly different topic, you've been specializing lately in big data projects. But I believe that goes beyond your teaching work and into using these tools to solve problems in the real world, more or less. Is that accurate? Yeah, so I am ADD. So in being independent, you know, I kind of like have a year of I'm going to teach, teach, teach, teach, teach for a year, because I want to go hang out in Santa Barbara, basically. And then I have a year of I'm going to teach, teach, teach, teach, teach for a year because I want to go hang out in Santa Barbara, basically. And then I have a year of, oh, I want to actually build something because, you know, if you don't actually build something, then you're not really the most effective teacher. Right. You're you're just kind of reading out of a book.

And plus, like anyone who's involved in technology, why we're involved is because it's a creative process and we want to actually build things. So I kind of go back and forth. Like for example, two years ago for 14 months straight, I was embedded with a dev team. So I would call myself an architect who codes. So it was a really interesting invite. It was an IoT project on Amazon for an enterprise. It was a sprinkler controller, literally. So we made the phone app for the sprinkler controller for golf courses and stuff like that. And we were one of the first people to go out on AWS IoT. And so it was just a super interesting engagement. And I literally was coding every day,

either in pair or in group. They call it mob programming at this particular company. And sitting there and seeing the pain when people are working with a new API, seeing the learning around working with new protocols, seeing the learning. This was their first public cloud project. So I have this combination of building actual things and then teaching about it. Fascinating. I've always more or less stayed away from data in general. My background was always in things that would these days be called stateless,

but we never thought of it that way once upon a time. But databases, data stores, data lakes, big data, data tributaries, data estuaries, whatever it is we call them, was always the stuff that scared the heck out of me. Because if you break the data in some form, it's very difficult, expensive, and sometimes impossible to get that back. So that's the stuff that leaves scars. As a result, I stayed away from that. And I don't have a whole lot of visibility into the quote-unquote rise of big data these days. What's driving the, I guess, VC frenzy, if nothing else, around the entire area of big data? Is it real or is it hype? Well, now I'm working, my production thing that I'm doing right now is in genomics. So

specifically helping to speed up the processing of the results of aggregate genomic sequencing, trying to find the needles in the haystack for disease conditions. I'm working with this group in Australia, which is called CSIRO, the Commonwealth of Scientific and Industrial Research Organization, which is basically the National Science Foundation of Australia. And they've developed some customized libraries for processing because the amount of data coming off the sequencers is of a volume we've really never seen before. So certain data collection situations and genomics, in my experience, is probably the lead driver of this because human DNA, for example, is 3 billion with a B data points for each sample. And when you are doing machine learning with looking for variants of interest, so different000, 3,000 samples, that's a matrix calculation space of 1.7 billion data points. So, yes, would be my short answer.

I'm doing it. data that big. But one of the reasons, in addition to the humanitarian aspect, which is helping to speed up research around personalized treatment for things like cancer, is just the intellectual challenge of true big data that I think is going to trickle into other domains. I see IoT data volumes rising, not to the extent of genomics, but exponentially based on what we've seen. Do you find that the rise of all this data at edge is necessitating the idea of pre-processing this data before you send it through? Or are companies still trying to shove it across a network and then do the heavy lifting once it arrives? You know, it's such a paradigm shift, right? You know, people still think of pre-processing, batching, and they don't go to the obvious thing, which is streaming.

And the reason is because streaming was seen as a specialty use case, and the skills around big data streaming are relatively hard to find. So you see a lot of old thinking, even to the extent of what some of the groups i've worked with in the cloud you know they're going to work with virtual machines or past solutions like elastic map reduce on amazon and we're really trying to take advantage of some of the newer technologies around containers and serverless for example we're trying to work on a prototype with SageMaker, which is containerized machine learning for one of the tools with the CSIRO group to get them off of servers. But everybody starts there, right? Because you can't just suddenly jump to all the new technologies. You have to start with the old patterns, and then you have to figure out what are the limits in

terms of time and cost, and when does it make sense to move to the newer technologies. To that end, tying back to what we said at the very beginning, how you have deep roots in all three of the large public cloud providers that we see in North America these days, all three of these companies, AWS, Azure, and GCP are coming out with a number of big data-oriented tools that are aimed at, I guess, removing the quote-unquote undifferentiated heavy lifting to steal Dr. Vogel's term and effectively start delivering insight and ability to separate signal from noise from all of these data offerings. At this point, from your perspective, is any particular vendor a clear leader of the pack at this moment, or are they all more or less still waiting for someone to break out? Well, I think what really started this, and they were actually, as they often are, ahead of the market, was Google BigQuery. I think the product was just, people just couldn't understand

because, you know, you open a query window and you upload some many, many, many text files and you just can do an ANSI SQL query and there's no servers and no nothing to manage. And they just, they, it was just magical. They couldn't understand it. It's interesting that Amazon waited till last year to bring out Athena because, you know, they could have done it sooner. But Amazon is very good at engaging the customer market. When is the customer ready for the new transitions? And I think that's one of the reasons why they are so dominant in so many markets. Because they certainly have fantastic technologies, but they have the appropriate balance for many situations, especially bringing out new paradigms

of understanding when the market can accept the new paradigm. And the uptick on Athena has been pretty substantial. So they seem to have gotten that one right. Now, the other one that I'm looking at is Glue. And I think that Glue is a very elegant product, but I have difficulty getting customers to understand what's really going on there. So they might have been a little early with Glue. I'm not going to shame anyone because I don't think I understand what's going on with Glue. Can you distill it down for me? Sure, sure. It's extract, transform, and load, or pre-processing, as you would say, as a service. So rather than

setting up virtual machines and, you know, having a bunch of scripts and a bunch of batch processes to, let's say, denormalize or deduplicate or fix nulls or all that kind of cleaning stuff that you do with data, you have recipes in PySpark and you have, you know, the flexibility of containers, which is a lot cheaper and more scalable in terms of your ETL. And again, don't get me wrong. I personally think Glue is fantastic. And I'm, you know, being an educator, I'm not, doesn't break my heart that there's a need for education because that's, I work in that area too. But I'm just seeing with customers that they instantly understand Athena, SQL queries on files, they get that.

But in terms of ETL as a containerized service, and Amazon's the only one struggling here, by the way. Google has a great product called DataFlow, which is the productization of Apache Beam, which is similar conceptually. And then they even put another product on front of it, which is kind of a GUI interface, almost like a SQL Server Integration Services interface called Data Prep, which generates data flow code. But again, the uptake hasn't been, I think, what the companies are wanting because in the data world,

change is more slow than in the programming world. And so the DBAs and the people that are used to working with the licensed products moving to container-based cloud services, that's a pretty big jump. It seems like a common theme where people will take a look at a new offering from AWS, GCP, et cetera, and say that it's garbage because it lacks a certain feature, it has a certain failure mode that doesn't work for a particular use case, and then they'll write it off and tend to dismiss it. And I think that that's a relatively naive approach. I see the promise of a lot of the services that have been coming out over the past couple of years. And are they ready for prime time now? Probably not. And some of these companies would do well to call out some of these shortcomings before people trip over them.

But what excites and inflames my passions is the ability to see where this is going, or to at least have a sense of, okay, like take Lambda, for example. Yes, picture a version of Lambda that doesn't have the current limitations of it as implemented today, and you start to see the world unlock in a number of different ways. I imagine a lot of this is true for data with the added caveat of if something working on your data screws up the first time, you're probably not likely to give it another try three months later just because you've been so badly burned the first time. Is that accurate or is that me bringing my I touched a database once and now I'm not allowed within 500 yards of one ever again bias speaking? Well, before you start a new project, I call phase zero data hygiene, which is when's the last time that you did a trial restore? Not do you have a backup? When is the last time you successfully did a trial

restore? I'll have to ask my DBA about that. There you go. And if you don't have that, and sometimes phase zero takes a year. Because if you have, okay, I'm telling you a real world story. I had one company, they had big data, they copied their data nine times. I mean, that's one way to do it. Just, you know, if it doesn't work for some tool, just make another copy. But, you know, I mean, clearly that's not what we want to do. So you have to have your house in order before you put this new stuff on top. And, you know, that's what you work with an independent consultant for rather than working directly with a vendor.

Because the vendor just wants you to get to use their product. Whereas somebody who's had a little bit of real world experience says, okay, before we go on this new stuff, let's make sure that we're having clean data put in because it's still garbage in, garbage out, right? Absolutely. And I'm privileged in that my data exploration projects in my own consulting work is generally limited to Amazon bills. And so we're talking on the high side in tens of gigabytes, which at that scale is, okay, I'm going to save a copy of that before I start doing anything too ridiculous to it because disk space is cheap. And I'm not manipulating it directly in place. I'm still dealing with the size of data that on some machines can fit in RAM and fits on any laptop that I've used in recent memory. So that doesn't generally tend to hold up at petabyte scale. Yeah, well, the whole thing with moving towards data lakes, too, that encourages experimentation.

Because storage in S3 or Google Cloud Storage or Azure Blobs is phenomenally cheaper than storing in RDS or Redshift or Spanner or whatever. And that whole move. And so that's one of those hygiene things I'm trying to help customers with. You know, you still run, you know, even though we're talking about like all this fancy serverless stuff, you still run into people that have never archived their data. Like they have 10 years of data on a relational database because they just did, right? So, you know, as you're moving them to the cloud, you go, all right, let's partition off the data you're not using, which tends to be 75%

of the database, throw that into S3, and then you can experiment with that because it's really cheap, right? So it's that discipline that you want to have before you really get going. The other thing that I want to mention is, because this is something I've really seen come up in the last year, and it's really given customers value quickly, is the machine learning APIs. You know, there's so much hype right now around TensorFlow and MXNet and all the other ones, which they're great for their particular purpose. But I just actually worked on a course on machine learning,

and I spent more time because of that course working with the Amazon machine learning APIs like Recognition and Lex and Polly and all those. And I am really, really impressed what is available now in terms of special purpose machine learning. And I'm going to be directing my customers to look at that before they go and build custom models. And I really think that combining that with getting historical data into S3, that's going to be opening up a lot of doors for customers to do interesting experiments. And it's really just kind of a tip I wanted to pass on because it's something I see very newly available really

across all three cloud vendors. A lot of competition there because of course whoever gets your data processing dollars is probably going to be the most successful cloud vendor. That's why storage prices are going to the bottom because they want you to store your data up on their data lake and then they're going to make their money on the processing, both ANSI SQL and increasingly machine learning. I'm seeing that in my own practice too, in that when you start looking at multi-cloud workloads or arbitraging between different providers, it works really well to say that I can move this containerized workload to a different

provider and save 20 cents an hour. The counter argument too, is that due to the data gravity and data transfer costs, migrating the data to a place where that container can work on it intelligently is a $20,000 project. So that tends to wind up restricting a lot of the compute to where the data lives. Very much so. Very, very much so. You know, I know you do work around cloud costs, and so do I.

In fact, I made a course around it because I had so many people, you know, asking me. And as you well know, understanding, particularly, and I have to call them out for this, Amazon cloud costs is almost a full-time job, unfortunately. And one of the things that I do hope happens is I do hope that the other vendors push Amazon towards better tooling. I have to celebrate Google because if you're, for example, spinning up a VM there, it's literally a slider. And you put a slider of how many CPUs, how much space you want, and you can see it shows you the price on the page. And I would really, you know, people will say, what do you want from Amazon? And I've said this over and over and over.

I want the slider. I don't want to know about 57 different EC2 instance names and numbers. I want the slider. I want simplification. But I think the other vendors are going to drive that. GCP is definitely ahead of the pack with respect to costing approaches. The idea of kill all billable resources in a project is phenomenal. I've been saying for a while that my entire business, which is fixing the horrifying AWS bill, should not exist. I would love a day where Amazon releases a product or a service

that renders me completely irrelevant. And I get to go work on other problems in this space that aren't the bill. I shouldn't have to effectively be a finance department's data science team more or less, and set up these convoluted processes and controls. What I do shouldn't be a thing, but we're somehow in this weird place where I don't see a clean way for AWS to fix this in the near term. Yeah, I think you have deeper insight than I do.

But again, I've literally been hired to do just that, similar to you. Like I had one startup, they were spending $200,000 a month on getting no return. I said, oh, I'm going to take a percentage. I can just fix this. I wish I found more people willing to be okay with percentages. I would be able to retire in a month. So we've talked a little bit about education, and we've talked a little bit about the rise of various cloud

services in the big data space. Let's put those two together. As cloud proliferates further and further, how does that impact technical education? So it's interesting. My work in this area follows my own daughter. I have one kid, and she's now in university. And she actually started early. She went to Stanford in the summer when she was 17 because she's a good student. And that combined with some other work I'd done with universities. I was a judge at a data hackathon last year in Southern California, 15 teams. You have to tell you this is a true story. 15 teams in the Expedia data set and of the 15 teams, zero, zero used the cloud. Most of them subsampled the data. The winning team split the data set across their seven laptops so they could process it.

And I looked at them and I said, oh, my goodness, have you guys never heard of EC2? And they're like, yeah, but we just don't know. So I have this vision and I've been shopping out to the cloud vendors of creating a series called Cloud for College Students. And I actually have a set of students lined up worldwide. And so I have it on my agenda. Now that I'm going to talk about it, I'm going to have to do it. I want to do it this year and I want to have a public series. And I've surveyed the students to see what their pain is because, you know, they're only going to be interested if they have pain, right? So I'm going to start with something really basic, like, have you ever lost homework? Well, how do you do a multi-cloud redundant backup? Because they've

all lost homework, right? And then another thing, have you ever had to install something on your laptop and it screwed up your Python build because you didn't know how to do virtual environments? Well, here's how to do a virtual machine. Or for some of the data science students, have you ever had to wait two days to run your workload because you had an old laptop? Well, here's how to get an Amazon machine learning AMI with GPUs on it. So, I mean, I've got the whole vision. So I'm going to start educating. You know, I mean, it starts with one person. We need a lot more, but I think that we

need to start with our college students because these kids are not using the cloud, which is, I mean, seriously, they split the data on seven laptops and these are the top universities in California. Are you kidding me? Something that I found is that, as you mentioned earlier, you're afflicted with some more professional ADD. I have actual ADD as well. And one of the challenges I have is I have difficulty paying attention to certain forms of learning. For me, sitting in a class, I'll zone out. Make me watch a video, I have to keep rewinding because my mind wanders and I miss something important that was just said. Reading is the way that I tend to absorb things most

effectively. So I fall into the trap sometimes of assuming that this is how everyone learns. I had to be convinced to start a podcast because I'm generally not a podcast listener for some of those reasons. The challenge that I find is that most people that I interact with are not like me, thank the Lord. But they tend to learn better from someone teaching them through guided learning. And it seems that a lot of the online courseware that I'm seeing doesn't have the same interactive element where you have a teacher there to ask questions to. Online forums are full of people referencing, I saw one last night, Zed Shaw's Learn Python the Hard Way. Someone was asking if the people in the forum could take them through

the program that he used as an example and explain it line by line because they didn't get it. And to some extent, that strikes me as the exact sort of thing you should be able to ask a teacher about. But it feels like we're moving more and more towards unidirectional, one-to-many learning styles. Is that just a selection bias on my side, or is this something that is starting to look like a paradigm shift? It is, and it's concerning to me. Obviously, I have a bias because I have been a teacher of adults for so many years in classrooms, both public classrooms and in private company engagements, and I've taught synchronously, both live and online for, you know, gosh, 15 years. So I, you know, I just come right up front and state my bias that teachers are an important part of the process.

But in addition to my bias, I've done some research, you know, I've read people's research work around the impact of multimodal teaching, as you talked about, and that's something a skilled teacher will bring to a classroom. So the simplest way, auditory, visual, and kinesthetic. But I mean, there's a lot more nuances to it. So I believe passionately that although the cloud is ready for serverless, here's our soundbite, education is not ready for teacherless. I don't think it's a good thing. I think you just titled the episode. I'm continuing to get data because opinions are only that. But as I continue to gather data, this is something that I think I'm going to continue to speak and write about because I think that it is an important change in our society, one that we need to understand the implications of.

And with everything else, we need to have data around. I have a personal bias towards meeting people where they are. Everyone learns differently and being able to convey information in a way that resonates with the audience in a way they can digest has always been important. It's why I think that there isn't only one way to do things. And I fall into the trap semi-frequently of assuming that everyone is going to learn the way that I am. It turns out that most people don't generally tend to learn best by reading man pages in the dark in their parents' basement, but that was my approach to it. A common theme among conversations that I've had with various guests on this podcast is starting to emerge. And that is the road that many of us walked to get to the point of technical skill that we have today

isn't there anymore. I don't see the help desk jobs that existed in the same way that did when I held those roles. I don't see email sysadmin wanted as a job description that exists. And having used AWS for just shy of 10 years now, even today, I log into the console, I see the fine print, and oh my word, that's not the fine print, that's the list of services. I have lost sight of what it would take to get to where we are have either done so by absorbing such a body of knowledge that it's not reasonable to expect people to pick that up, or worse, that those of us who have gotten here to some extent have pulled the ladder up behind us. very little more than an interest and a natural aptitude for absorbing information to start becoming a modern technologist in today's world? Well, again, I am focusing now on college students because I have one, and so I have access to the student community basically um and the other situation is i'm uh continuing to implement my 20 or 25 volunteer rule of life so the project i'm doing in australia i'm actually not being paid

for um and over my many years of being a technical professional, I've done three substantial projects. One with an electronic medical records project in Zambia for five years, 10 years of teaching kids programming, and now I'm in, this will be my second year of working with CSIRO. And so I personally do this 20% time. When I was employed, I made it a condition, part of my contract, so it's paid and bonused for the DigiGirls work. And I engage with newer technology people in this work. So I will work with, you know, homemakers who are transitioning into the workforce. I'll work with students. I'll work with people who are just coming into technology. And sometimes they'll work with me for free. Sometimes I will pay them at a reduced rate. So again, I can't solve the problem at scale, but I can just say that what is working

for me is I'm kind of doing like apprenticeships. I have one student now, he was one of the winners of the hackathon that worked for me for 40 hours for free. And now he's an apprentice and we're working on refactoring some of the Scala code for the random forest for this variant spark genomic algorithm. So I'm taking on personal apprentices. I don't know, again, that's not very scalable, but that's what I'm doing. Yeah. It's fantastic seeing people taking apprentices, because I do feel this is to some extent, something of a trade in that you learn by doing. What I'm trying to figure out, and I'm sure I'm not alone in this one, is how you start scaling that. But even now, if I were to take on an

assistant and start brain dumping to them all the stuff that I've picked up about AWS alone, it feels like there would still be a year or two of foundational knowledge that I've just either forgotten or taken for granted. So just getting people from a pedagogical standpoint to a point where they can have those conversations and start to contribute in a meaningful way is a difficult place to get to. I wish I saw a clearer path there. Oh, I do too. I mean, I wish there was some sort of, you know, national, I don't know, support or program or something kind of like a job core. I don't know. I mean, I'm not a politician, but it just seems to me we have this huge gap. I mean, again,

I took some of the students from the data hackathon to a one-day event in L.A. that was hiring. And they were all graduating seniors and they all had CS degrees or data science degrees or both. And some of them got offers for free internships and stuff. But I was astounded that because they didn't have experience with the cloud, and it was usually Amazon, but basically they were asked any cloud, and none of them had any experience with the cloud. None of them got any job offers. And, I mean, that was, again, one of the reasons why I want to do, I took four students, and I took them around to all my vendor friends, and nobody got a job offer because they didn't have cloud experience. And I thought, well, I have to do something about this.

Entry-level role, three years experience required. Yep. I wish there was a clearer path. We almost need Lynn as a service to some extent, but that gets dangerously close to teacherless. Well, maybe I can make a bot with Lex, right? The Lynn bot. Have it pop up in Slack and correct you from time to time when you're about to say something foolish. That'd be great. I love that model. It's kind of annoying to me, but I feel that way about every chat bot that I tend to encounter. So is there anything you'd like to mention or leave us with as we close out? Something you want us to remember or think about as we wait the long

week until the next episode? Yeah. I mean, I'm a personal example of lifelong learning, and I'm really trying to bust the ageism myth. You live up in the Valley, so good for you. I only visit occasionally, but I am over 50, and I am, you know, doing TensorFlow and doing the hottest, newest thing and learning, you know, relearning linear algebra. And, you know, my brain is not broken, people of the computing world. So, you know, I always hold up myself as an example. And then I often have people who are in a similar age group come up to me and say, thank you for saying that because it needs to be said. You know, it's a real problem in our industry that when you be really, it's I think 40, when you get to be 40, you're considered to be a manager or useless, which is basically sort of the same thing. And I am I am hardcore technical and I am over 50 and I am going strong.

And I'd like to see more people kind of alongside. Don't be afraid to try to learn something new. You're going to have to study, but you can study. Brain's not broken. You know, This has increasingly been top of mind. I am creeping up on 40 in the next few years. And that is something that I'm starting to see that we have almost a cult here in the Valley that worships youth, where if you're not a 22-year-old willing to work 110 hours a week on your startup, then you don't have what it takes to succeed. And that's a toxic, painful myth. Speaking with you is always a delight because I don't believe we've ever had a conversation where I didn't come away brimming with new ideas and inspiration and things to research that you've just touched on briefly. You don't get that level of depth and breadth by having gone to a bootcamp

for 18 weeks. This is something that you learn by doing, and there's definite value in experience. So thank you for saying that. It's something that I don't think we see touched upon enough these days. You're welcome. Thank you very much for joining me. This was Lynn Langit.

I'm Corey Quinn, and this is Screaming in the Cloud. This has been this week's episode of Screaming in the Cloud. You can also find week's episode of Screaming in the Cloud. You can also find more Corey at screaminginthecloud.com or wherever fine snark is sold.